TL;DR
Lenovo announced at ISC that the ongoing memory shortage, dubbed ‘RAMageddon,’ is expected to last for years due to AI demand and industry shifts. The company shared a ‘survival guide’ for navigating high memory costs, highlighting fundamental changes in the memory market that may persist into the late 2020s.
Lenovo has declared that the ‘RAMageddon’ — the ongoing memory shortage and price surge — will become the new normal, with no immediate relief expected. The company shared this outlook during a presentation at the International Supercomputing Conference (ISC), emphasizing that memory prices are unlikely to return to previous lows soon, due to sustained demand from AI infrastructure and industry changes.
According to reports from ComputerBase, Lenovo’s presentation included a slide titled ‘The 5 Step RAMageddon Survival Guide,’ indicating a recognition of the prolonged memory crunch. Lenovo’s broader message is that the economics of the memory industry have fundamentally shifted, with demand from AI and data centers absorbing new manufacturing capacity coming online around 2028. The company pointed out that even as capacity expands, prices are unlikely to drop significantly because of persistent high demand.
Lenovo cited SK hynix’s plans to triple its memory production capacity by 2034 as supporting evidence that the industry is not expecting a return to the oversupply and low prices seen in early 2025. The firm also highlighted that designing servers now requires considering higher memory capacity, with upcoming servers featuring 16 memory channels per processor and potentially requiring around 1 TB of RAM to fully utilize bandwidth. Industry forecasts from Micron and SK hynix align with Lenovo’s outlook, projecting supply constraints until at least 2027-2030, driven by AI infrastructure investments and long-term supply agreements worth hundreds of billions of dollars.
Additionally, the shortage has made high-margin memory types like HBM more attractive, as manufacturers shift production capacity toward high-value products for AI accelerators. This shift narrows the price gap between HBM and traditional DRAM, influencing how hyperscalers and enterprise users approach system design and deployment, often favoring GPU-attached HBM to reduce overall system memory costs.
Implications of the Prolonged Memory Shortage for Industry
The announcement underscores that high memory prices and shortages are expected to persist into the late 2020s, impacting server design, infrastructure costs, and AI deployment strategies. As memory becomes more expensive and demand remains high, organizations may need to rethink hardware architectures, potentially increasing reliance on GPU-based memory solutions and optimizing memory utilization to manage costs. This shift could influence the broader technology ecosystem, affecting pricing, supply chains, and the pace of AI infrastructure expansion.
As an affiliate, we earn on qualifying purchases.
Industry Trends and Industry-Wide Memory Market Changes
The ongoing ‘RAMpocalypse’ has been characterized by a global memory shortage and rising prices across DRAM and NAND components, driven largely by surging AI demand and constrained supply. Companies like Micron and SK hynix have forecasted supply constraints lasting until 2027 or beyond, with long-term supply agreements indicating that hyperscalers are prioritizing securing memory capacity. Historically, memory markets have been cyclical, but recent industry shifts—such as increased manufacturing capacity and higher-margin product focus—suggest that the current tight supply may extend well into the late 2020s, marking a fundamental change in industry dynamics.
“It will never be like it was last year.”
— Lenovo presentation at ISC
A-Tech 128GB Kit (4x32GB) DDR5 5600MHz PC5-44800 ECC UDIMM 2Rx8 (EC4 9×4) Dual Rank 1.1V ECC Unbuffered DIMM 288-Pin Server, Workstation RAM Memory Upgrade Modules
- Compatibility: For select DDR5 servers and workstations only
- Total Capacity: 128GB kit with 4x32GB modules
- Module Type: DDR5 ECC unbuffered UDIMM
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Uncertainties Surrounding Memory Market Recovery Timeline
While industry leaders predict a prolonged shortage, the precise timeline for a market correction remains unclear. Factors such as technological advances, geopolitical influences, and potential shifts in AI demand could accelerate or delay supply normalization. Additionally, it is not yet confirmed how significantly new capacity additions will impact prices, given the high demand and industry restructuring.
The HBM Shock : What is the Memory Hegemony that Dominates the GPU Era (Japanese Edition)
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Next Steps for Industry and Technology Buyers
Memory vendors will likely continue prioritizing high-margin products like HBM for AI and enterprise applications. Industry observers should monitor capacity expansions, pricing trends, and supply agreements, as these will influence hardware costs and deployment strategies. Organizations deploying large-scale infrastructure may need to adapt their procurement plans and architecture designs to account for sustained high memory prices and potential supply constraints into the late 2020s.
CORSAIR Vengeance LPX DDR4 RAM 32GB (2x16GB) Up to 3200MHz CL16-20-20-38 1.35V Intel XMP AMD EXPO Computer Memory – Black (CMK32GX4M2E3200C16)
- Maximum Speed Note: Requires overclocking and BIOS adjustments
- Memory Chips: Hand-sorted for high performance
- Compatibility: Optimized for Intel and AMD DDR4 motherboards
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Key Questions
Will memory prices ever return to pre-2025 levels?
Industry forecasts suggest that prices may not revert to previous lows until late 2020s or beyond, due to persistent demand from AI infrastructure and industry restructuring.
How will the memory shortage impact server and data center costs?
Higher memory prices and capacity constraints are likely to increase overall infrastructure costs, prompting a shift toward more memory-efficient architectures and GPU-based solutions.
What strategies can organizations use to cope with ‘RAMageddon’?
Organizations may need to optimize memory utilization, consider alternative architectures like GPU-attached HBM, and plan procurement well in advance to mitigate supply risks.
Is this shortage specific to certain memory types?
The shortage affects DRAM and NAND components broadly, with high-margin products like HBM gaining relative attractiveness due to supply shifts and demand.
